Machine-learning based generation of process models from natural language text process descriptions is severely restrained by a lack of datasets. This lack of data can be attributed to, among other things, an absence of proper tool assistance for dataset creation, resulting in high workloads and inferior data quality. We address these shortcomings with a tool for annotating textual process descriptions. Compared to other, existing data annotation tools, ours implements a multi-step workflow specifically designed for extracting process information, including supporting features that have been shown to reduce workloads and improve data quality.